A database of children with autism spectrum disorder (ASD) living in the north east of England was established in 2003 to gather reliable information about numbers and needs. The database contains information about the child’s diagnosis, school, language level, behaviour, other medical diagnoses and estimated intellectual ability and is used for service planning and research.

BACKGROUND

Disability registers

The UK Children Act (1989)1 requires local authorities to hold a register of disabled children. Registration is voluntary and parental consent is required. The guidance, reiterated in the 1999 report Quality protects,2 recommends that local authorities should collaborate with health services to share information. A search uncovered no studies in the social science literature on local authority registers. However, Colver et al3 found in one local authority that the social services register recorded 5% of the children on the child health disability register and only 20% of the most severely disabled children. A UK Department for Education and Skills research report4 and the 2006 parliamentary hearings on services for disabled children5 again identified the need for better quality disability registers.

Most disability registers either cover a single condition or attempt to cover all known individuals in a defined population.6 Newton and Garner7 point out that the former have rich and accurate data but may be unrepresentative; the latter offer prevalence data, but obtaining consent from all children and families is difficult.

Current legal framework on health registers

There is general public support for the use of data for research and public health surveillance,8 but issues of consent are paramount. A European directive (1995)9 on the protection of personal data was followed in the UK by the Caldicott Committee report (1997),10 11 the Data Protection Act (1998) and the Health and Social Care Act (2001). Registers which hold non-anonymous information on individuals now require explicit individual consent (with some exceptions under section 60 of the Health and Social Care Act).

Prevalence of autism spectrum disorders

Reported median prevalence rates of autism have increased steadily (4.4 per 10 000 in studies published in 1966–1991, 12.7 per 10 000 in 1992–2001).12 Systematic review13 suggests much of the variation can be attributed to changing diagnostic criteria.

Recent studies report the prevalence rates of broader ASD are around 60 per 10 000.13^–19 This rate has been generally accepted for children up to the age of 6 years. Where older children are included, rates of 90 and 116 per 10 00020 21 have been reported.

In the current climate there are many challenges for a register of children with ASD. These include how best to engage with parents/carers and professionals, how to accurately identify all individuals with ASD in a defined area, and how to confirm consistent usage of diagnostic criteria based on high quality professional assessments.22 The National Autism Plan for Children (NAP-C)23 in the UK (and a similar process in the USA24) has set guidelines for identification, assessment, diagnosis and intervention for children with ASD.

Proposal for a prospective database of children

The north east of England has a number of local advantages which provide an opportunity to create a unique database:

• a relatively stable population of 1.4 million in six local authorities in Tyne and Wear and in Northumberland,25

• implementation of the recommendations of the NAP-C by local multi-agency teams,

• strong regional child and adolescent mental health and learning disabilities services, with training links to local teams in the use of ASD-specific diagnostic instruments,

• a regional special educational needs network, and a regional office of the National Autistic Society,

• local epidemiological expertise in child health and disability registers in Newcastle University (eg, in childhood cancer, cerebral palsy).26 27

Aims of Daslⁿe

A database of children with autism spectrum disorder living in the north east of England (Daslⁿe) was established in 2003 to gather reliable information about children with ASD across six authorities. It has the following aims:

• to assist health authorities, local authorities, health trusts and voluntary agencies to plan services by making available regular reports on the number of children with ASD at different ages and the broad needs of these children and their families as measured by indicators of severity;

• to allow analysis of anonymised data to provide information on changes in prevalence, severity and location;

• to enable families and voluntary organisations to be more involved in planning and service provision through access to information via regular reports and newsletters;

• to be a sampling frame for research studies. Such research could include aetiological and genetic studies, evaluation of interventions, and detailed study of the impact of ASD on family life and of support from services.

METHODS

History

The database evolved from meetings of representatives of health, education and social services, voluntary organisations and clinical researchers from Newcastle University in 2000. After the aims were agreed, ethics approvals and health trust permissions were obtained. A local charitable foundation provided development funds for 3 years from 2003 for the initial survey phase. Local authorities and health authorities agreed to fund ongoing shared administration costs.

Content of the database

Once parents have given informed consent, they complete the short Daslⁿe questionnaire (child’s diagnosis, school, language level, behaviour and other problems). Parents also nominate a professional to complete a questionnaire (based on ICD-10 diagnostic criteria for autism), which also provides information about other medical diagnoses and estimated intellectual ability.

For the initial survey we aimed to recruit all children up to the age of 12 years with a diagnosis of ASD so that useful population data would be available quickly. Subsequently, all parents of newly diagnosed children are approached. Where children are first registered before 5 years of age, information is updated at 6 years. No further updating occurs at present.

Structure

Daslⁿe is held on a stand-alone computer in Newcastle University. This ensures confidentiality and the database is not affected by health service or local authority restructuring. The Director is a senior clinical academic, working with an administrator and database designer. Daslⁿe is supported by four committees, each having parent representation. The Oversight Committee, chaired by a representative from the strategic health authority, includes senior service managers and monitors the project in relation to four criteria: comprehensiveness of coverage, levels of parental consent, consistency of diagnostic practices, and functioning local coordinating groups (see below). The Steering Group includes professionals from each local authority area and Newcastle University staff and takes policy decisions, shares experiences between areas, and receives reports from subcommittees. The Parent Advisory Group has played a crucial role: for example, parents advised that the word “database” was preferable to “register” because of potential confusion with the child protection register, they revised drafts of the information sheet and consent form, and suggested that parents should generally not be approached about Daslⁿe until 3 months after diagnosis. The Reports, Publications and Research Committee develops policies on how data can be requested by planners, parents and professionals, and how researchers can ask for help with recruitment to studies.

The other Daslⁿe structures are the six coordinating groups. In each local authority area, professionals from health and education services work together to maintain a complete list of the names of children with ASD in the survey age range. Ethics committee opinion allowed sharing of identifying information between services. Each family is given information about Daslⁿe by a professional known to them, in person or by post. Consent forms and questionnaires are returned directly to Daslⁿe. Monthly reports are provided to the local coordinating groups.

Publicity

Daslⁿe was launched in October 2003 with wide media coverage. A strong public identity has been created with a logo incorporating the Millennium bridge over the River Tyne, promotional materials, 6-monthly newsletters including articles on themes arising from the data, and a website (http://www.ncl.ac.uk/daslne). The first Daslⁿe conference in 2006 endorsed the database, celebrated its achievements and reflected on lessons learned.

RESULTS

Numbers

The provisional lists of children with ASD compiled by local professionals contained 986 children born after 31 December 1992, in line with predicted rates.13^–19 It has taken 3 years (as of October 2006) to approach all families in the initial survey, with two authorities only recently commencing recruitment. The response so far has been variable (table 1), with 491 consents and only 20 refusals (overall 51.8% response).

Several strategies have been used to increase participation, with success dependent upon the time and effort of the local coordinating group. For example, further approaches to parents were made by telephone in area 4 (table 1) resulting in a high response rate.

Data validation

Validation was examined under two headings:

1. Accuracy of diagnosis. Forty children were selected at random. For 20, information in their medical notes was checked against questionnaires. The remaining 20 were seen in person: the Autism Diagnostic Observation Schedule28 was administered by a research associate and parents completed the Social Communication Questionnaire.29 These checks confirmed that all children met criteria for autism or ASD, or had this diagnosis documented in their medical notes.

2. Completeness of ascertainment. This was estimated by capture/recapture analysis30^–32 in area 3 (table 1). In the initial survey, there were 343 notifications of children with ASD from six sources. Removal of duplications and those without a confirmed diagnosis left 189 children whose families were sent Daslⁿe information packs. As 11 more had already contacted Daslⁿe, there were 200 children in all. The lists may not be completely independent, but each agency recorded children separately. Using the three largest lists in two-list analysis, the number of predicted cases of ASD was between 221 and 224. The completeness of ascertainment was therefore high at 89–90%.

Representativeness

Anonymised lists provided by each local coordinating group were used to compare responders with non-responders by gender and Townsend deprivation index.33 Gender (6:1, boys:girls) did not differ significantly (χ2 = 0.28, p = 0.60). Social deprivation was significantly greater in non-responders (t = 5.61, p<0.001); however, the difference was small (mean 1.77 vs 0.45 for non-responders vs responders, on a scale from 10 (very deprived) to −6). Other surveys and registers show similar bias.7

Use of data

The data have been requested by various bodies, including the strategic health authority working group on services for ASD, enabling it to mount a successful bid to support the commissioning of adult services. Data also informed a feasibility study for a North of Tyne autism centre for children with ASD with the most challenging needs.

DISCUSSION

A disability register has four principal uses34: service planning, epidemiological research, clinical audit and individual patient care. This paper describes the setting up of a database of children with ASD to primarily address the first two aims, although local organisations can use Daslⁿe for the third.

The development of Daslⁿe has been more difficult than we expected. Obtaining parental permission has taken longer than anticipated in all areas. However, analyses suggest that Daslⁿe data are sufficiently representative to be usefully reported for service planning and to be used as a sampling frame for research and local authority consultation with parents. Already data are being used where regional comparisons are important.

The processes followed in establishing Daslⁿe were informed by Newton and Garner’s7 specifications for successful registers, including management by a multidisciplinary team, establishment of a steering group, focused aims, clear leadership, careful location of the database, and strong publicity. It is evident that success depends on keen leadership of the local coordinating groups; problems have arisen where this was lacking due to illness or change of staff.

Coordination of information sources for Daslⁿe has brought agencies together to improve local service provision; two have used the data to obtain funding for a coordinator for a multi-agency pathway for assessment of ASD. The partnership with parents in setting up Daslⁿe has been a considerable strength, enhancing credibility when new parents are approached. Such public awareness benefits are not shared by other approaches to establishing the prevalence of ASD, such as surveillance systems carried out through the inspection of records.19 35 However, the undertaking has been more complex than anticipated and may not encourage replication in other parts of the UK.

The main challenges now are to:

• establish stable funding from local sources and research bodies,

• ensure the data inform transition planning for young people; the older survey phase children will become adults in 2011,

• increase coverage in all the local authorities, and establish consistency in involving families of newly diagnosed children,

• ensure the resource is utilised fully by local authorities, health bodies and voluntary organisations.